On Thu, May 14, 2015 at 11:44:54AM +0200, nudge wrote:
> I've slightly reworded your text for clarity:
Thanks, with guidance from the chairs, I'll merge as much as I can.
> This section defines a protocol for avoiding deprecated digest
> algorithms when these are published in a peer's TLSA RRset
> along-side stronger algorithms. A mixture of algorithms MAY be
> present in server TLSA records to allow for interoperability with
> legacy or constrained clients.
I don't think that the "MAY" above needs to be in RFC 2119 language.
I had a lower case "may", in order to discuss a hypothetical. This
is not a paragraph that *specifies* allowed server behaviour.
However, if "MAY" really is better, it does not make a great deal
of difference.
> If a server's TLSA RRset contains RRs with more than one digest
> matching type and adheres to the requirements of <xref target="rrreq"/>
> with each combination of TLSA parameters containing at least
> one record that matches the server's current certificate chain
> (or raw public keys) then under these circumstances, a client MAY
> identify its preferred digest algorithm and only process records
> for that algorithm, in addition to any records with matching type
> Full(0).
Here (and this is not the result of your changes) we run into a
bit of trouble. Because of the original uncertain status of the
consensus on agility, this says "MAY", but a couple of paragraphs
the text says that the client SHOULD employ algorithm agility. We
need to pick one or the other for both paragraphs.
> To allow for digest algorithm agility, all published DANE TLSA RRsets
> MUST conform to the requirements of <xref target="rrreq"/>. With
> servers publishing compliant TLSA RRsets, TLS clients MAY, for each
> combination of usage and selector, ignore all digest-based RRs except
> those that employ the strongest digest algorithm. The client then
> processes only those RRs plus any records with matching type Full(0)
Ditto.
> TLS clients SHOULD use digest algorithm agility when processing
> the DANE TLSA records of a TLS server. Any algorithm agility MUST
> be applied after first discarding any unusable or malformed records
> (unsupported digest algorithm, or incorrect digest length). Thus,
> for each usage and selector, the client SHOULD process only usable
> records whose digest algorithm is considered to be the strongest,
> as well as any records with a matching type of Full(0).
The above is the final "SHOULD" that is in apparent conflict with
the initial "MAYS" that explain potential strategies, which are
then upgraded from potential to "SHOULD".
So the exposition needs to lose some hedging, and state clearly
whether Digest Agility is "MAY" or "SHOULD" throughout.
--
Viktor.
_______________________________________________
dane mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dane
